Ten percent polarized optical emission from GRB 090102
arXiv:1010.1255v1 [astro-ph.HE] 6 Oct 2010
I.A. Steele1 , C.G. Mundell1, R.J. Smith1 , S. Kobayashi1, C. Guidorzi2
1
Astrophysics Research Institute, Liverpool John Moores University, CH41 1LD, U.K.
2
Physics Department, University of Ferrara, Via Saragat 1, 44122 Ferrara, Italy
The nature of the jets and the role of magnetic fields in gamma-ray bursts (GRB) remains unclear1, 2 . In a baryon-dominated jet only weak, tangled fields generated in situ through shocks would be present3 . In an alternative model, jets are threaded with large scale magnetic fields that originate at the central engine and which accelerate and collimate the jets4 . The way to distinguish between the models is to measure the degree of polarization in early-time emission, however previous claims of γ-ray polarization have been controversial5–8 . Here we report that the early optical emission from GRB 090102 was polarized at the level of P = 10 ± 1%, indicating the presence of large-scale fields originating in the expanding fireball. If the degree of polarization and its position angle were variable on timescales shorter than our 60-s exposure, then the peak polarization may have been larger than 10 per cent. The standard GRB fireball model3 comprises an initial compact emitting region, expanding relativistically, in which internal shocks dissipate the bulk energy, converting kinetic to radiated energy, the so-called prompt emission. As the shell of the relativistically expanding fireball collides with the surrounding circumburst medium, a forward shock is produced, which propagates outwards through the external medium and results in the long-lived afterglow whose emission is detectable from X-ray to optical, infrared and, in some cases, radio wavelengths. Interaction of 1
the relativistic fireball with the ambient medium also produces a short-lived reverse shock that propagates backwards through the expanding shell9, 10 .
Exploiting the ability of robotic optical telescopes to respond rapidly and automatically to the discovery of new GRBs, a custom, fast-response, optical polarimeter11 (RINGO) was deployed on the 2.0 meter robotic Liverpool Telescope12 (La Palma) with the goal of measuring the degree of polarization of optical emission from GRBs at early time. RINGO uses a rotating Polaroid to modulate the incoming beam, followed by corotating deviating optics that transfer each star image into a ring that is recorded on a CCD (Figure 1). Any polarization signal present in the incoming light is mapped out around the ring in a sin2θ pattern. RINGO was first used in 2006, when it observed GRB 060418 at 203s after the gamma ray burst and coincident with the time of deceleration of the fireball. At this time the reverse (assuming it was present) and forward shock components would have contributed equally to the observed light. For GRB 060418 a 2σ upper limit on optical polarization of P7×105 s post-burst14. The absence of any additional emission components from late-time central engine activity superimposed on the afterglow light curve allows a straightforward interpretation of the light curves in the context of current GRB models. The steep-shallow decay of optical emission from GRB 090102 is characteristic of an afterglow whose early-time light is dominated by fading radiation generated in the reverse shock9, 18 .
Figure 1 shows the RINGO exposure obtained on the night of 2009 January 2. The afterglow of GRB 090102 is clearly visible, as are six brighter foreground objects. Detection of these objects allowed us to perform additional checks on the instrumental calibration at the time of the GRB. In addition by observing the same region of sky at later dates after the GRB had faded (2009 January 28, 2009 April 18 and 2009 May 19) the stability of RINGO was also verified. The ˚ polarization of GRB 090102 is P = 10.2 ± 1.3%, in contrast to measured optical (4600 - 7200A) the foreground objects which show P ∼ 1−4% (Figure 2). A simple Monte Carlo analysis (Figure 3) was performed to estimate the significance of the polarization measurements. This showed that the rank of our GRB measurement amongst a distribution of randomly reordered GRB trace data was 9,988/10,000.
3
In interpreting our measurement first we consider whether such a polarization could be produced via the production of magnetic instabilities in the shock front (Figure 4(c)). A very optimistic estimate of the coherence length can be made by assuming it grows at about the speed of light in the local fluid frame after the field is generated at the shock front - in this situation polarized radiation would come from a number of independent ordered magnetic field patches. A measured polarization of 10% is at the very uppermost bound for such a model19 and therefore seems unlikely. As an alternative to the “patch” model, we have also considered the case where the observer’s line of sight is close to the jet edge20 (Figure 4(b)). In this case since the magnetic fields parallel and perpendicular to the shock front could have significantly different averaged strengths21 a polarization signal can also be produced. However applying this model to GRB 090102 we would have expected a steepening of the light curve (a “jet-break”) just after the time of our polarization measurement rather than the observed flattening. Similarly there is no evidence of a jet break in the X-ray light curve up to late times. Furthermore, our detection of 10% is much higher than the reported polarization signal of a few % associated with a jet break in late time afterglow of other events22, 23 . We also rule out an Inverse Compton origin for the optical polarization - a mechanism suggested to explain earlier γ-ray polarization measurements24 - in which lower energy photons are scattered to higher energies by colliding with electrons in the relativistic flow. If Inverse Compton emission is present, it is more likely to contribute to the high-energy X-ray and γ-ray bands than the optical band and again requires the observer’s line of sight to be close to the edge of the jet (Figure 4(b)) to produce significant polarization which as we have already discussed is not the case for GRB 090102.
4
It therefore seems apparent that in the case of GRB 090102 the high polarization signal requires the presence of large-scale ordered magnetic fields in the relativistic outflow (Figure 4(a)) . As the measurement was obtained while the reverse-shock emission was dominant in GRB 090102, the detection of significant polarization provides the first direct evidence that such magnetic fields are present when significant reverse shock emission is produced. Magnetization of the outflow can be expressed as a ratio of magnetic to kinetic energy flux, σ. The degree of magnetization cannot be sufficient for the jet to be completely Poynting flux dominated (σ > 1) since this would be expected to suppress a reverse shock25 . We can therefore reconcile the detection of polarization in GRB 090102 and our previous non-detection in GRB 060418 in a unified manner if GRB jets have magnetization of σ ∼ 1. In the GRB 060418 case, the jet would have had slightly higher magnetization than unity, resulting in the suppression of a reverse shock, while GRB 090102 would have σ slightly smaller than unity, which is optimal to produce bright reverse shock emission. Of course due to the small sample (only two objects), we can not rule out a possibility that each GRB jet had very different magnetization.
Finally we note that a high degree of polarization is also predicted for the prompt γ-ray emission in the presence of large-scale ordered magnetic fields26, 27 . Recent claims of rapidly (∼ 10 s) variable γ-ray polarization from less than 4% to 43% (±25%) in the prompt emission of GRB 041219A28 lend further support to models with magnetized outflows and offer the possibility that the peak optical polarization from GRB 0901012 could have been even higher than that measured in our 60 second exposure.
5
1. Lyutikov, M. The Electromagnetic Model of Gamma Ray Bursts. New J. Phys., 8, 119-143 (2006) 2. Zhang, B. & M´esz´aros, P. Gamma-Ray Bursts: Progress, Problems & Prospects. Int. J. Mod. Phys. A, 19, 2385-2471 (2004) 3. Piran, T. Gamma Ray Bursts and the Fireball Model. Phys. Rep., 314, 575-667 (1999) 4. Lyutikov, M., Magnetocentrifugal launching of jets from discs around Kerr black holes, Mon. Not. R. Ast. Soc., 396, 1545-1552 (2009) 5. Coburn W., Boggs S.E., Polarization of the prompt γ-ray emission from the γ-ray burst of 6 December 2002, Nat.,. 423, 415-417 (2003) 6. Rutledge R.E, Fox, D.B., Re-analysis of polarization in the γ-ray flux of GRB 021206, Mon. Not. R. Ast. Soc., 350, 1288-1300 (2004) 7. Willis D.R., et al., Evidence of polarisation in the prompt gamma-ray emission from GRB 930131 and GRB 960924, Astron. Astrophys., 439, 245-253 (2005) 8. McGlynn et al. 2007, Polarisation studies of the prompt gamma-ray emission from GRB 041219a using the spectrometer aboard INTEGRAL, Astron. Astrophys., 466, 895-904 (2007) 9. Zhang, B., Kobayashi, S., M´esz´aros, P., Gamma-Ray Burst Early Optical Afterglows: Implications for the Initial Lorentz Factor and the Central Engine, Astrophys. J., 595, 950-954 (2003) 10. Zhang, B. & Kobayashi, S. Gamma-ray Burst Early Afterglows: Reverse Shock Emission from Arbitrarily Magnetized Ejecta. Astrophys. J., 628, 315-334 (2005) 6
11. Steele, I.A., et al., RINGO: a novel ring polarimeter for rapid GRB followup, Proc. SPIE, 6269, 179S (2006) 12. Steele, I.A., et al., The Liverpool Telescope; performance and first results, Proc SPIE, 5489, 679-692 (2004) 13. Mundell, C.G., et al. Early Optical Polarization of a Gamma Ray Burst Afterglow Science, 315, 1822-1824 (2007) 14. Mangano, V. et al. Swift Observations of GRB 090102, GCN-Report-192.1 (2009) 15. Klotz, A. et al. GRB 090102: TAROT Calern Observatory Optical Observations. GCN Circ. #8761 (2009) 16. Covino, S. et al. GRB 090102: REM Observations of a Bright Afterglow. GCN Circ. #8763 (2009) 17. Gendre B., et al., Testing the GRB models with the strange afterglow of GRB090102, Mon. Not. R. Ast. Soc., submitted (2009). 18. Gomboc, A. et al. Multiwavelength Analysis of the Intriguing GRB 061126: the Reverse Shock Scenario and Magnetization. Astrophys.J., 660, 489-495 (2008) 19. Gruzinov, A. & Waxman, E. Gamma-Ray Burst Afterglow: Polarization and Analytical Light Curves. Astrophys. J., 511, 852-861 (1999) 20. Gruzinov, A. Strongly polarized Optical Afterglows of Gamma-Ray Bursts. Astrophys. J., 525, L29-L31 (1999) 7
21. Medvedev, M.V., Loeb, A., Generation of Magnetic Fields in the Relativistic Shock of Gamma-Ray Burst Sources, Astrophys. J.,, 526, 697-706. (1999) 22. Covino S. et al, GRB 990510: linearly polarized radiation from a fireball, Astron. Astrophys., 348, L1-L4 (1999) 23. Wijers, R.A.M., et al., Detection of Polarization in the Afterglow of GRB 990510 with the ESO Very Large Telescope, Astrophys. J., 523, L33-L36 (1999) 24. Lazzati, D., Rossi, E., Ghisellini, G., Rees, M. J., Compton drag as a mechanism for very high linear polarization in gamma-ray bursts, Mon. Not. R. Ast. Soc, 347, L1-L5 (2004) 25. Mimica, P., Giannios, D., Aloy, M.A., Deceleration of arbitrarily magnetized GRB ejecta: the complete evolution, Astron. Astrophys., 494, 879-890 (2009) 26. Fan, Y., Interpretation and implications of the non-detection of GeV spectrum excess by the Fermi Gamma-ray Space Telescope in most gamma-ray bursts, Mon. Not. R. Ast. Soc., 397, 1539-1548 (2009) 27. Granot J,. The Most Probable Cause for the High Gamma-Ray Polarization in GRB 021206, Astrophys. J., 596, L17-21 (2003) ˇ Variable Polarization Measured in the 28. G¨otz D., Laurent, P., Lebrun F. Daigne,. Boˇsnjak, Z., Prompt Emission of GRB 041219A Using IBIS on Board INTEGRAL, Astrophys. J., 695, L208-L212 (2009)
8
29. Schmidt, G. D., Elston, R., Lupie, O. L., The Hubble Space Telescope Northern-Hemisphere grid of stellar polarimetric standards, Astron. J., 104, 1563-1567 (1992) 30. Clarke, D., Neumayer D, Experiments with a novel CCD stellar polarimeter, Astron. Astrophys., 383, 360-366 (2002).
Acknowledgements The Liverpool Telescope is operated on the island of La Palma by Liverpool John Moores University in the Spanish Observatorio del Roque de los Muchachos of the Instituto de Astrofisica de Canarias with financial support from the UK Science and Technology Facilities Council. CGM acknowledges financial support from the Royal Society and Research Councils U.K. We thank Dr Jon Marchant for preparing Figure 4 for us.
Author Contributions IAS: Instrument design and build. Data reduction and analysis. CGM: Instrument science case, scientific interpretation. RJS: Data reduction and analysis. SK: Theoretical interpretation. CG: Coordination of observations and identification of afterglow.
Competing Interests The authors declare that they have no competing financial interests.
Correspondence Correspondence and requests for materials should be addressed to IAS.
(email:
[email protected])
9
Figure 1. RINGO observation of the field of GRB 090102 observed 2009 Jan 2. The field of view is 4.6 x 4.6 arcmin. The data have been dark subtracted and flat-fielded using standard astronomical algorithms. The afterglow of GRB 090102 is labeled G along with six foreground sources (labeled 1–6). Foreground source 5 is contaminated by an overlapping faint source, and so was not used in further analysis. We followed our standard RINGO reduction procedure in which flux traces for all objects on all nights were extracted within annuli with inner (8 arcsec) and outer (14 arcsec) radii sufficient to ensure that seeing variations do not influence the extracted fluxes. The traces were then sky subtracted by the normalized flux inside the inner trace radius and
10
divided through by an average of the traces from routinely obtained zero-polarization standards29 to remove the known 2.7% instrumental polarization. The resulting flux traces for a sample of objects and the GRB are presented in Figure 2.
11
Figure 2. RINGO data for GRB 090102 and calibration sources. Example flux traces around the rings of three of the foreground objects (1-3) and GRB 090102 showing a clear sin 2θ signal for the GRB are shown. For the foreground objects traces are presented taken simultaneous with GRB 090102 (filled symbols) and on the night of 2009 May 19 (unfilled symbols). Analysis30 of different subsets13 of the data in the GRB trace allows a mean polarization and standard deviation to be measured, giving a value of 10.1±1.3% for GRB 090102. Objects 2 and 3 have low polarization (≤ 1.5%) in both exposures and set limits on uncorrected instrumental polarization effects. Object 1 is detected as weakly polarized (2.5%) in both measurements, demonstrating the stability of the instrumental setup (the instrument reference position angle varies due to the telescope altitude-
12
azimuth mount between the two epochs, so the traces are not in phase, however the amplitude of variation and hence derived polarization is similar). Objects 4 and 6 (not plotted) show similar stable weak polarization between different epochs of ∼ 3% and ∼ 4% respectively.
13
Figure 3. Monte Carlo simulation using GRB090102 data. The distribution of measured polarizations derived from randomly reordered GRB trace data demonstrates the significance of the actual result. The measured value for the GRB (10.1%) is shown with an arrow, and can be seen to be highly significant (rank = 9,988/10,000). Similar analyses for the foreground objects in the frames confirms that objects 2 (rank 780/1,000) and 3 (rank 540/1,000) have no detectable polarization at the level of 1.5% and that objects 1 (2.5% - rank 969/1,000) , 4 (3.3% - rank 927/1,000) and 6 (4.1% - rank 913/1,000) have measured polarizations in line with the expected values for stars within our galaxy29.
14
Figure 4. Competing models of GRB magnetic field structure. The schematic shows three representations of a GRB outflow in the context of the standard fireball model for a variety of magnetic field structures and different orientations to the observer’s line of sight (optical axis). A large degree of polarization is predicted when the ejected material is threaded with a largescale ordered magnetic field as shown in (a) and is the favoured model to explain the measured polarization in GRB 090102. Alternatively, if no ordered magnetic field is present and, instead a tangled magnetic field is produced in the shock front, the detected light will be polarized only if the observer’s line of sight is close to the jet edge (b). In this case, however, the predicted steepening of the light curve that is expected when observing an off-axis jet is inconsistent with the flattening exhibited in the light curve of GRB 090102. A compromise is shown in (c) in which the shock
15
front contains a number of independent patches of locally-ordered magnetic fields; a measured polarization of 10% is at the very uppermost bound for such a model.
16
